Contributed talk in Robot Control 1, July 29, 2019, 11:30 a.m. in room USB.4.005

Different Forms of Random Motor Activity Scaffold the Formation of Different Habits in a Simulated Robot

Mario Zarco, Matthew Egbert

watch Publication The concept of habit as a precarious self-sustaining pattern of sensorimotor behavior has added the notion of plasticity to the concept of sensorimotor autonomy. One can imagine how some habits would be better suited to the maintenance of a biological individual. Evolution can bias the parameters of the plastic medium over which sensorimotor autonomy emerge so as to be conducive to biological autonomy. In this work, we show that modulating some parameters that bring about plastic changes in the behavior-generating medium, different sensorimotor individuals emerge. We present how different random motor activity biases habit formation when a sensorimotor-habit-based controller is coupled with a simple robot in a one-dimensional environment. The simulation demonstrates that, varying the parameters of a random-based exploratory phase activated by the controller when the robot is in sensorimotor conditions it has not experienced previously, qualitative different habits emerge that are characterized by static, monotonic and oscillatory behaviors. The experiment proves what kinds of habits are more likely to be performed by the robot given specific values of the parameters that modulate random motor activity. Frequencies distribution of the oscillatory behaviors is also presented so as to understand what kind of quantitative features can be encouraged.